Results

cesm2.ssp245

diff_of_means ratio_of_sd amplitude_ratio_of_means maximum_error ks_mean_on_coarse_res_with_extremes qqplot_mae acf_mae extremogram_mae
lstm.cesm2.ssp245 0.11% 0.986 0.932 0.348 0.132 0.333 0.035 0.021
cnn.cesm2.ssp245 0.20% 1.001 0.935 0.120 0.161 0.590 0.059 0.039
nv.cesm2.ssp245 -0.51% 0.997 0.956 0.126 0.137 1.482 0.034 0.041
xgboost.cesm2.ssp245 -0.53% 0.947 0.818 0.256 0.165 1.541 0.080 0.019

Time series of the first days

How Often Peaks Hit Hourly

QQ Plot

Distribution of the undownscaled value on days with estimated extremes values.

On the x-axis we have the daily mean (standardized). It says Undownscaled value, but is the daily mean after the downscaling. A good idea is to plot the original undownscaled value.

The purpose of this plot is to illustrate the distribution of P(undownscaled value | we predicted an extreme). This is useful because it reveals how much information we can recover concerning extreme events. If the distribution is skewed to the right, it suggests that we’re predicting extreme values only when extreme values have already occurred. Conversely, if the lower tail of the distribution resembles the reanalysis data, it indicates that we can capture short-duration extremes (e.g., brief periods of heavy rainfall, such as an intense downpour lasting an hour before stopping).

Autocorrelogram

Extremogram

cesm2.ssp370

diff_of_means ratio_of_sd amplitude_ratio_of_means maximum_error ks_mean_on_coarse_res_with_extremes qqplot_mae acf_mae extremogram_mae
lstm.cesm2.ssp370 0.12% 0.982 0.929 0.368 0.069 0.353 0.033 0.009
cnn.cesm2.ssp370 0.21% 0.995 0.928 0.122 0.058 0.609 0.058 0.016
nv.cesm2.ssp370 -0.57% 0.967 0.959 0.127 0.041 1.651 0.019 0.017
xgboost.cesm2.ssp370 -0.59% 0.917 0.809 0.247 0.118 1.735 0.072 0.017

Time series of the first days

How Often Peaks Hit Hourly

QQ Plot

Distribution of the undownscaled value on days with estimated extremes values.

On the x-axis we have the daily mean (standardized). It says Undownscaled value, but is the daily mean after the downscaling. A good idea is to plot the original undownscaled value.

The purpose of this plot is to illustrate the distribution of P(undownscaled value | we predicted an extreme). This is useful because it reveals how much information we can recover concerning extreme events. If the distribution is skewed to the right, it suggests that we’re predicting extreme values only when extreme values have already occurred. Conversely, if the lower tail of the distribution resembles the reanalysis data, it indicates that we can capture short-duration extremes (e.g., brief periods of heavy rainfall, such as an intense downpour lasting an hour before stopping).

Autocorrelogram

Extremogram

cesm2.ssp585

diff_of_means ratio_of_sd amplitude_ratio_of_means maximum_error ks_mean_on_coarse_res_with_extremes qqplot_mae acf_mae extremogram_mae
lstm.cesm2.ssp585 0.19% 0.979 0.928 0.357 0.056 0.552 0.033 0.011
cnn.cesm2.ssp585 0.28% 0.995 0.927 0.124 0.062 0.818 0.059 0.019
nv.cesm2.ssp585 -0.44% 0.995 0.958 0.111 0.053 1.294 0.031 0.016
xgboost.cesm2.ssp585 -0.47% 0.947 0.811 0.249 0.089 1.379 0.082 0.012

Time series of the first days

How Often Peaks Hit Hourly

QQ Plot

Distribution of the undownscaled value on days with estimated extremes values.

On the x-axis we have the daily mean (standardized). It says Undownscaled value, but is the daily mean after the downscaling. A good idea is to plot the original undownscaled value.

The purpose of this plot is to illustrate the distribution of P(undownscaled value | we predicted an extreme). This is useful because it reveals how much information we can recover concerning extreme events. If the distribution is skewed to the right, it suggests that we’re predicting extreme values only when extreme values have already occurred. Conversely, if the lower tail of the distribution resembles the reanalysis data, it indicates that we can capture short-duration extremes (e.g., brief periods of heavy rainfall, such as an intense downpour lasting an hour before stopping).

Autocorrelogram

Extremogram

ec_earth3.ssp434

diff_of_means ratio_of_sd amplitude_ratio_of_means maximum_error ks_mean_on_coarse_res_with_extremes qqplot_mae acf_mae extremogram_mae
lstm.ec_earth3.ssp434 0.15% 0.999 0.947 0.378 0.146 0.425 0.048 0.021
cnn.ec_earth3.ssp434 0.23% 1.016 0.983 0.138 0.122 0.677 0.056 0.022
nv.ec_earth3.ssp434 -0.27% 1.019 0.946 0.163 0.176 0.775 0.065 0.021
xgboost.ec_earth3.ssp434 -0.28% 0.965 0.718 0.259 0.285 0.838 0.139 0.032

Time series of the first days

How Often Peaks Hit Hourly

QQ Plot

Distribution of the undownscaled value on days with estimated extremes values.

On the x-axis we have the daily mean (standardized). It says Undownscaled value, but is the daily mean after the downscaling. A good idea is to plot the original undownscaled value.

The purpose of this plot is to illustrate the distribution of P(undownscaled value | we predicted an extreme). This is useful because it reveals how much information we can recover concerning extreme events. If the distribution is skewed to the right, it suggests that we’re predicting extreme values only when extreme values have already occurred. Conversely, if the lower tail of the distribution resembles the reanalysis data, it indicates that we can capture short-duration extremes (e.g., brief periods of heavy rainfall, such as an intense downpour lasting an hour before stopping).

Autocorrelogram

Extremogram

mri_esm2_0.ssp245

diff_of_means ratio_of_sd amplitude_ratio_of_means maximum_error ks_mean_on_coarse_res_with_extremes qqplot_mae acf_mae extremogram_mae
lstm.mri_esm2_0.ssp245 0.16% 0.978 0.917 0.383 0.195 0.504 0.043 0.029
cnn.mri_esm2_0.ssp245 0.25% 0.987 0.922 0.136 0.189 0.737 0.062 0.039
nv.mri_esm2_0.ssp245 -0.32% 1.153 0.962 0.112 0.304 1.031 0.086 0.100
xgboost.mri_esm2_0.ssp245 -0.34% 1.094 0.922 0.216 0.294 0.983 0.097 0.020

Time series of the first days

How Often Peaks Hit Hourly

QQ Plot

Distribution of the undownscaled value on days with estimated extremes values.

On the x-axis we have the daily mean (standardized). It says Undownscaled value, but is the daily mean after the downscaling. A good idea is to plot the original undownscaled value.

The purpose of this plot is to illustrate the distribution of P(undownscaled value | we predicted an extreme). This is useful because it reveals how much information we can recover concerning extreme events. If the distribution is skewed to the right, it suggests that we’re predicting extreme values only when extreme values have already occurred. Conversely, if the lower tail of the distribution resembles the reanalysis data, it indicates that we can capture short-duration extremes (e.g., brief periods of heavy rainfall, such as an intense downpour lasting an hour before stopping).

Autocorrelogram

Extremogram

mri_esm2_0.ssp370

diff_of_means ratio_of_sd amplitude_ratio_of_means maximum_error ks_mean_on_coarse_res_with_extremes qqplot_mae acf_mae extremogram_mae
nv.mri_esm2_0.ssp370 -0.13% 1.109 0.961 0.114 0.287 0.631 0.068 0.034
xgboost.mri_esm2_0.ssp370 -0.17% 1.061 0.884 0.224 0.299 0.541 0.096 0.025
lstm.mri_esm2_0.ssp370 0.18% 0.978 0.925 0.365 0.230 0.538 0.032 0.016
cnn.mri_esm2_0.ssp370 0.26% 0.989 0.935 0.131 0.243 0.766 0.050 0.025

Time series of the first days

How Often Peaks Hit Hourly

QQ Plot

Distribution of the undownscaled value on days with estimated extremes values.

On the x-axis we have the daily mean (standardized). It says Undownscaled value, but is the daily mean after the downscaling. A good idea is to plot the original undownscaled value.

The purpose of this plot is to illustrate the distribution of P(undownscaled value | we predicted an extreme). This is useful because it reveals how much information we can recover concerning extreme events. If the distribution is skewed to the right, it suggests that we’re predicting extreme values only when extreme values have already occurred. Conversely, if the lower tail of the distribution resembles the reanalysis data, it indicates that we can capture short-duration extremes (e.g., brief periods of heavy rainfall, such as an intense downpour lasting an hour before stopping).

Autocorrelogram

Extremogram

mri_esm2_0.ssp434

diff_of_means ratio_of_sd amplitude_ratio_of_means maximum_error ks_mean_on_coarse_res_with_extremes qqplot_mae acf_mae extremogram_mae
lstm.mri_esm2_0.ssp434 0.15% 0.988 0.922 0.381 0.203 0.431 0.043 0.012
nv.mri_esm2_0.ssp434 -0.17% 1.155 0.963 0.108 0.276 0.878 0.085 0.053
xgboost.mri_esm2_0.ssp434 -0.20% 1.099 0.901 0.215 0.297 0.724 0.106 0.020
cnn.mri_esm2_0.ssp434 0.24% 0.999 0.925 0.128 0.230 0.688 0.063 0.023

Time series of the first days

How Often Peaks Hit Hourly

QQ Plot

Distribution of the undownscaled value on days with estimated extremes values.

On the x-axis we have the daily mean (standardized). It says Undownscaled value, but is the daily mean after the downscaling. A good idea is to plot the original undownscaled value.

The purpose of this plot is to illustrate the distribution of P(undownscaled value | we predicted an extreme). This is useful because it reveals how much information we can recover concerning extreme events. If the distribution is skewed to the right, it suggests that we’re predicting extreme values only when extreme values have already occurred. Conversely, if the lower tail of the distribution resembles the reanalysis data, it indicates that we can capture short-duration extremes (e.g., brief periods of heavy rainfall, such as an intense downpour lasting an hour before stopping).

Autocorrelogram

Extremogram